Many whole body oncology studies using positron emission tomography (PET) are performed without correcting for the effects of attenuation. Whether this improves or degrades the detection of tumors has been the subject of much discussion. This project was undertaken to determine the effect of attenuation correction (AC) on the detectability of tumors in PET studies using 18-FDG. The method used employed a novel technique of Ogated replicatesO which permitted actual measurments of noise in patient studies, with and without AC. By comparing the intensity of the signal from tumors to the noise in the tumor region and to the noise in surrounding background regions, it was found that expected tumor detectability would be optimum using AC in the central portions of the body, while using no AC was optimum only in peripheral regions of the body, and in portions of the lung. If the AC were combined with an iterative reconstruction algorithm, however, the AC method was superior to the no-AC method at all locations of the body studied. The preliminary results of the study were accepted for oral presentation, and a full manuscript is under preparation.